The implementation of tertiary treatment at large wastewater treatment plants (WWTP) may be required in many WWTPs in Germany due to water quality targets defined in the Water Framework Directive (EU-WFD) and Bathing Water Directive (EU-BWD) of the European Union. Furthermore, potential environmental risks of organic micropollutants (OMP) from anthropogenic sources (i.a. pharmaceuticals, sweeteners) could require additional treatment steps for tertiary treatment in future. EU-WFD requires a “good ecological status” of all water bodies, which can lead to a need of enhanced phosphorus removal at large WWTP (>100’000 pe), targeting an effluent quality <100µg/L TP. Moreover, if a WWTP discharges upstream of bathing water, EU-BWD requirements have to be met. Hence implementing a disinfection step might be necessary. Different options for enhanced P-removal and disinfection have already been analyzed in their economic and environmental impacts (KWB 2013). Based on these results, both targets can be adequately met by coagulation with subsequent dual media filtration (DMF) and UV-disinfection (UV). On this basis, the present study focusses on the additional integration of a process for OMP-removal into a tertiary treatment scheme. Considered technologies for OMP-removal are oxidation by ozonation and adsorption by activated carbon (AC) either by dosing powdered activated carbon (PAC) or using filtration units with granulated activated carbon (GAC), respectively. These technologies increase the additional demand of energy and chemicals for tertiary wastewater treatment. WWTPs are already one of the major contributors of electricity demand at municipality level (UBA 2008), and further treatment steps may add up significantly in this environmental impact. In the present study, different options and process configurations for OMP-removal are integrated in a tertiary treatment with advanced P-removal and UV-disinfection, and the entire tertiary treatment train is then analysed in its environmental impacts using the methodology of Life Cycle Assessment (LCA). The goal of the LCA is to reveal the trade-off between local environmental benefits by higher effluent quality and global environmental impacts, e.g. an increasing CO2-footprint. With the methodology of LCA different tertiary treatment schemes are analysed in a holistic approach “from cradle to grave” (ISO 2006), which includes direct effects at water bodies through discharge, and indirect effects resulting from infrastructure, chemical and electricity demand by tertiary treatment and additional sludge treatment. The baseline scenario is defined as treatment of secondary effluent of an existing WWTP located in Berlin, Germany (1’500’000 pe) by DMF with coagulation and UV (Figure 1.1). Sludge from backwash of filtration units is considered in the LCA by a simplified model for sludge treatment and mono-incineration (SMIP). For integration of OMP-removal into tertiary treatment, 7 possible scenarios are compared in their environmental impacts (Figure 1.2): (1) Ozone+DMF+UV, (2) PAC-dosing+DMF+UV, (3) PAC-cycle+DMF+UV, (4) DMF+GAC-filter+UV, (5) DMF w/ GAC-layer+UV, (6) Ozone+DMF w/ GAC-layer+UV, or (7) parallel treatment by ozonation and PAC+DMF+UV, respectively. Each scenario is analysed with a low, medium, and high dosage of ozone or AC, displaying the whole range of economic feasibility and effluent quality targets (Table 1.1). The specific dosage of ozone and PAC are referred to DOC-concentration of the secondary effluent (12.8mg/l DOC). Data used for advanced P-removal and UV-disinfection are based on a previous study (Remy et al. 2014) using planning data from the WWTP operator considering process efficiency, infrastructure, energy and chemical demand. Data for OMP-removal technology are based on pilot plants and planning data from WWTP operator. For LCA, impact categories of ReCiPe Midpoint method are taken into account (Goedkopp et al. 2008), e.g. global warming potential (GWP) or freshwater eutrophication potential (FEP), and cumulative energy demand (CED) of fossil and nuclear resources (VDI 2012), and USEtox indicators (Rosenbaum et al. 2008) freshwater ecotoxicity (ETP) and human toxicity potential (HTP). Environmental benefits of tertiary treatment scenarios on the global scale can be seen in the FEP and ETP indicators. TP from secondary effluent is reduced from 320µg/l to 55µg/l TP after tertiary treatment. The global USEtox indicator ETP includes preliminary impact factors for seven measured OMPs (6 pharmaceuticals, 1 herbicide), neglecting potential toxic effects of metabolites or transformation products as limitation of the multi-fate model. Removal of OMP has a positive effect on ETP in all scenarios. However, background processes and heavy metal loads play a major role in the contribution to the global ecotoxicity indicator. On the contrary, a higher energy and chemical consumption lead to a significant increase of CED and GWP due to OMP-removal (Figure 1.3). Comparing baseline scenario (DMF+UV) with the gross GWP of a large WWTP, the CO2-footprint will increase by +11% (82g CO2-eq/m³). Ozonation increases the GWP by 23% to 37% depending on ozone dosage. Main contributors for GWP are electricity and liquid oxygen demand for ozonation. Highest effects on GWP are detected for the scenario “PAC-cycle+DMF+UV” with an additional CO2-footprint of 36% or 110%, respectively, which is mainly caused by emissions during production of AC. In summary, OMP-removal can double the GWP of an existing large WWTP in the worst case and thus contributes significantly to global environmental effects. Production of AC is a crucial parameter for scenarios using GAC or PAC. Hence, a sensitivity analysis is performed changing raw materials for AC production. AC production is modelled according to available data from Bayer et al. (2005) using 3kg of hard coal as resource for activation process and generating 1kg of virgin AC. Other possible resources for AC production can be lignite or coconut shells. Varying the type of resource reveals a high uncertainty in GWP. Considering scenario “PAC+DMF+UV” a possible reduction of -23% of net GWP using coconut shells or even an increase of net GWP by +32% using lignite is possible. Since specific discharge limits for OMP removal are not defined yet, a direct comparison between the considered scenarios is not possible, as they lead to different effluent qualities in OMP concentration. Thus, in theory a low dosage of PAC (1.0g/gDOC) may be sufficient to achieve certain effluent targets, whereas ozonation could require a high dosage (1.0g/gDOC) for the same quality, or vice versa.

Im Rahmen des Projekte OgRe wurde das Ausmaß der Belastung von Regenablauf für Berlin durch ein einjähriges Monitoringprogramm in Regenwasserabfluss der Trennkanalisation unterschiedlicher Einzugsgebietstypen (Altbau, Neubau, Gewerbe, Einfamilienhäuser, Straßenablauf) untersucht. Ziel war, eine möglichst vollständige Erfassung organischer Spurenstoffe zu erreichen (einschließlich Identifizierung zusätzlicher Substanzen durch non-target-Analytik). Darüber hinaus sollte geklärt werden, inwieweit die unterschiedlichen Einzugsgebietstypen ein unterschiedliches Spektrum an Belastung durch Spurenstoffe aufweisen. Diese Informationen wurden dann genutzt, um eine Hochrechnung der über das Regenwasser in die Gewässer gelangenden Spurenstofffrachten für Gesamt-Berlin und einzelne Gewässerabschnitte zu ermöglichen. Die erhaltenen Frachten wurden verglichen mit modellierten Frachten abwasserbürtiger Spurenstoffe, die über Kläranlagenablauf in die Berliner Gewässer gelangen. Insgesamt wurden etwa 90 volumenproportionale Mischproben auf ein Set von etwa 100 Spurenstoffen analysiert. Zusätzlich wurden 12 Regenereignisse in der Panke beprobt, um Spitzenkonzentrationen regenwasserbürtiger Spurenstoffe im Gewässer zu ermitteln und ins Verhältnis zur Trockenwetterbelastung (5 Proben) zu setzen. Auch eine Untersuchung mikrobiologischer Parameter und der zeitlichen Dynamik konnten im Rahmen des Projektes durchgeführt werden.

Managed aquifer recharge (MAR) is a widely accepted method for augmenting water supplies for potable and non-potable use. The success of the MAR system is often defined by a substantial removal of chemical and biological contaminants during subsurface passage. To determine removal rates and to differentiate between removal and overall attenuation due to dilution, estimation of mixing proportions is a key element of tracer applications. This report provides an overview of tracers suitable for MAR and discusses advantages and disadvantages for each tracer. The ideal tracer may be defined by: a natural or anthropogenic origin, a clear uneven distribution in the studied system (e.g. sharp contrast between source and native groundwater), non-toxicity (human and environmental), easy and cost-effective measurement, and a conservative (neither sorbed nor (bio-)chemical reactive) or at least predictable chemical or physical behavior. A huge number of tracers exist, each with advantages and disadvantage. Tracers can be dissolved (e.g. chloride, bromide), stable or radioactive isotopes (e.g. 18O, 3H), gaseous (e.g. SF6) or a physical properties (e.g. temperature). The use of heat as a tracer has several advantages over hydrochemical tracers. Temperature is inexpensive, easy and a robust parameter to measure. In contrast to chemical tracers, no laboratory analysis is required and the data is available immediately. Finally, a multi tracer approach (= 2 tracers) is always recommended, because the ideal tracer is rarely found. A reasonable combination is at least one conservative tracer (e.g. stable isotopes of water) with a retarded tracer (e.g. temperature) to evaluate short travel times from the point of recharge (e.g. riverbed or pond) to the recovery well.

The Australian Guidelines for Water Recycling – Managed Aquifer Recharge provide a ready-to -use and user-friendly compendium of knowledge. Practical instructions and checklists provide a step wise approach with a strong focus on implementation. The proposed models for water flow and substance transport allow a first tier estimation of present concentrations in ambient groundwater and the impacted zone in the aquifer. The use of stochastic models is not mandatory within the guidelines. A criticism which can be identified related to the use of models simply based on point estimates, is that especially in early stage risk assessments, where uncertainties are usually high, these models tend to pretend a level of certainty which often does not represent reality. Risks associated to inorganic chemicals are required to be treated with more detail. Rigorous quantification of biodegradation kinetics (e.g. first-order rate constants) and adsorption parameters (e.g. linear distribution coefficients) for EOCs during subsurface passage determined on field scale are still scarce. It is clear that first-order rate constants and linear distribution coefficients provide only a simplified description of the removal mechanisms during subsurface passage, because they neglect spatial and temporal dynamics of physical and chemical conditions. Nevertheless, this approach often provides a good approximation and allows also for site independent comparison of removal processes. Regarding the demonstration site in Berlin-Tegel the analysis showed that if the model of the Australian Guidelines is applied to the MAR system the travel time of 50d during subsurface passage cannot be guaranteed. In Germany, a residence time of 50d is usually considered to sufficiently reduce the risk of microbial hazards. Although risk calculations did not reveal immediate concern, it is recommended to develop and apply suitable verification monitoring techniques to quantify travel times and reduce present uncertainties. Moreover, this risk assessment and the study about the influence of the groundwater replenishment site on ambient groundwater (Sprenger and Grützmacher, 2015) clearly showed the need for protective measures against the input of undesired substances from shallow ambient groundwater.

Micropollutant concentrations found in stormwater runoff were extrapolated to annual loads at the scale of the city of Berlin (impervious connected area of ~170 km2). Extrapolation was done by city structure, i.e., it was assumed that concentration patterns found in one of five specific city structure types is representative for every area of this structure type. Preliminary results show that micropollutants of several substance types can enter Berlin surface waters at loads in the order of kg/yr via stormwater runoff: plasticizers (e.g., sum of Di-iso-decylphthalate and Di-iso-nonylphthalate at 770 kg/yr), flame retardants (e.g., tris(2-butoxyethyl) phosphate (TBEP) at 89 kg/yr), biocides from different sources (e.g., Glyphosate at 17 kg/yr and Mecoprop at 30 kg/yr), vulcanizing accelerator benzothiazole (as sum of benzothiazole and metabolites methylthiobenzothiazole and hydroxybenzothiazole at 65 kg/yr) and combustion byproduct polycyclic aromatic hydrocarbons PAH 16 (sum of 16 EPA PAH at 107 kg/yr). These loads are in a similar order of magnitude as micropollutants that enter Berlin surface waters via (treated) sewage, such as pharmaceutical residues carbamazepine and ibuprofen with estimated annual loads of 436 kg/yr and 35 kg/yr, respectively.

In recent years, the effect of urbanization on the quality of stormwater runoff gained increased attention including investigations on micropollutants. Especially in cities dominated by separated sewer systems, stormwater runoff containing micropollutants from anthropogenic origin is discharged mostly untreated into surface waters and therefore a potential source of high loads of pollutants. In a one year monitoring campaign stormwater runoff from five different catchments in Berlin is analyzed for major groups of micropollutants such as phthalates, organophosphates, organotin-compounds, biocides/pesticides, PAH’s, alkylphenols, polybrominated diphenylether, polychlorinated biphenyls and heavy metals. Sampling sites are equipped with automatic samplers, flow and water level meters in order to prepare flow proportional composite samples (recommended sampling strategy according to DIN ISO 5667-10). First results show that all groups of micropollutants were found in at least one catchment type in concentrations > 2 µg/L. Concentrations of the different micropollutant groups vary depending on the catchment types. So far, no organotin-compounds, polybrominated diphenylether or polychlorinated biphenyls were determined.

Untreated stormwater runoff can be an important source of pollutants affecting urban surface waters. For example, in Berlin each year 78% or 54 million m³ of stormwater are discharged mostly untreated into receiving surface waters. Beside “classic” stormwater pollutants (e.g. suspended solids, COD, phosphorous or heavy metals), trace organic substances such as biocides, plasticizers, flame retardants and traffic related micropollutants (e.g. vulcanizing accelerators originating from tire wear or combustion by-products such as PAHs) started to come into focus in recent years (Zgheib et al. 2012, Gasperi et al. 2014). To evaluate for the first time city-wide annual loads of these trace organic substances entering urban surface waters through stormwater discharge, an event-based, one-year monitoring program was set up in the city of Berlin.

A study is conducted to determine the relevance of micropollutants in urban stormwater runoff. To evaluate for the first time city-wide annual loads of stormwater-based micropollutants entering urban surface waters, an event-based, one-year monitoring program was set up in separate storm sewers in Berlin. Monitoring points were selected in 5 catchments of different urban structures (old building areas <1930, newer building areas >1950, single houses with gardens, roads >7500 vehicles/day and commercial areas) to consider catchment-specific differences. Storm events of different characteristics were sampled up to four hours during different seasons by automatic samplers triggered by flow meters. Volume-proportional samples (one composite sample per event) were analysed for a set of 100 parameters including 85 organic micropollutants (e.g. flame retardants, phthalates, pesticides/biocides, PAH), heavy metals and standard parameters. So far (70/88 samples), 60 organic micropollutants were at least once detected in stormwater runoff of the investigated catchment types. Concentrations were highest for phthalates with average concentrations of 13 µg/L for diisodecyl phthalate. For heavy metals, concentrations were highest for zinc (average: 950 µg/L). Results also showed catchment-specific differences for many compounds as well as seasonal differences for selected pollutants which can be used to improve micropollutant strategies and potentially prevent loads at the source.

We investigate water quality of a small urban river during dry and wet weather conditions, including both standard parameters and trace organics. The monitored river stretch receives both effluents from WWTP as well as (separate) stormwater runoff of an impervious area of 11 km2. Results show increases in concentrations in the river during rain events with a factor > 20 for zinc, polycyclic aromatic hydrocarbons, two herbicides and one flame retardant. Also, substances which are expected both in WWTP effluent and in stormwater effluents were detected at important concentrations in the river during wet weather, such as the corrosion inhibitor Benzotriazole (0.8 µg/L on average) and the plasticizer Diisodecyl phthalate (4.0 µg/L on average). The presented results are preliminary and will be complemented by more results and substances as well as an assessment of the relevance of the findings.

Stormwater treatment technologies to manage runoff during rain events are primarily designed to reduce flood risks, settle suspended solids and concurrently immobilise metals and nutrients. Life Cycle Assessment (LCA) is scarcely documented for stormwater systems despite their ubiquitous imple- mentation. LCA modelling quantified the environmental impacts associated with the materials, con- struction, transport, operation and maintenance of different stormwater treatment systems. A pre- fabricated concrete vortex unit, a sub-surface sandfilter and a raingarden, all sized to treat a func- tional unit of 35 m3 of stormwater runoff per event, were evaluated. Eighteen environmental mid-point metrics and three end-point ‘damage assessment’ metrics were quantified for each system's lifecycle. Climate change (kg CO2 eq.) dominated net environmental impacts, with smaller contributions from human toxicity (kg 1,4-DB eq.), particulate matter formation (kg PM10 eq.) and fossil depletion (kg oil eq.). The concrete unit had the highest environmental impact of which 45% was attributed to its maintenance while impacts from the sandfilters and raingardens were dominated by their bulky ma- terials (57%) and transport (57%), respectively. On-site infiltrative raingardens, a component of green infrastructure (GI), had the lowest environmental impacts because they incurred lower maintenance and did not have any concrete which is high in embodied CO2. Smaller sized raingardens affording the same level of stormwater treatment had the lowest overall impacts reinforcing the principle that using fewer resources reduces environmental impacts. LCA modelling can serve as a guiding tool for practitioners making environmentally sustainable solutions for stormwater treatment.